This proposal will demonstrate the effects of two housing models for homeless mentally ill adults and determine the relationship between client characteristics, client preferences, clinician judgment and client outcomes in these residences. The study population will consist of homeless individuals from Boston's mental health "transitional" shelters who suffer from severe and persistent mental illness. Sixty beds at each of two housing options will be available to study participants: independent living and specially designed group homes. The group homes will begin with 24-hour program staff but be designed to foster increased client participation in all aspects of program management with a gradual replacement of professional staff by resident consumers. Consumers experienced in self-help activities will provide on-site consultation and training. By the beginning of year 3 of the grant, it is expected that the residence will function as an autonomous, client-run home. Prior to random assignment to a housing option, we will assess the client's housing preference and the judgment of two clinicians regarding the best housing choice. In addition, each subject will be assigned an intensive clinical case manager, who will maintain ongoing contact with the client on at least a weekly basis and facilitate services provision. Subjects will be interviewed every six months to assess their current symptomatology, level of functioning, satisfaction with their living situation and changes in their housing preferences. Data will be collected on individual characteristics, frequency and intensity of contact with their case managers, use of other community-based mental health services, need for and use of physical health care providers, and psychiatric hospitalization. Subjects will be followed for eighteen months after entry into their housing alternative. The evolution of the consumer=run homes will be described.